Improved numerical procedure for the coupling of dynamic components using frequency response functions

Teik Chin Lim, Glen C. Steyer

Research output: Contribution to journalConference articlepeer-review

19 Scopus citations


Many dynamic systems of interest consist of coupled low and high modal density components. In these systems it is desirable to perform system simulation using component models characterized by a measured dynamic compliance matrix for the high modal density component and an analytical model for the low modal density one. Such a model requires inversion of the compliance matrix to obtain dynamic stiffness matrix for system assembly. This modeling approach has not been widely successful in the past due to measurement error magnification in the compliance matrix inversion. Further difficulties may be caused by compliance matrix rank deficiency and numerical error associated with extensive matrix algebraic manipulation. In this paper, a coupling procedure is proposed based on the singular value decomposition (SVD) theorem and the constrained mode component model which avoids these problems. The rigorous SVD routine performs a pseudo-inverse which also filters measurement noise and rank deficiency in the compliance matrix. The use of a constrained mode finite element model (FEM) eases parametric studies and leads to a simpler formulation. This procedure also allows for better representation of different component damping levels in the system simulation.

Original languageEnglish (US)
Pages (from-to)902-908
Number of pages7
JournalProceedings of the International Modal Analysis Conference - IMAC
StatePublished - 1991
Externally publishedYes
EventProceedings of the 9th International Modal Analysis Conference Part 2 (of 2) - Florence, Italy
Duration: Apr 15 1991Apr 18 1991

All Science Journal Classification (ASJC) codes

  • General Engineering


Dive into the research topics of 'Improved numerical procedure for the coupling of dynamic components using frequency response functions'. Together they form a unique fingerprint.

Cite this